When students first learn calculus, integration is usually presented as the process of finding the area under a curve.

For many functions, this approach works beautifully. We divide the x-axis into small pieces, form rectangles, add their areas, and then let the rectangles become infinitely thin.

This is the idea behind the Riemann integral.

For a long time, mathematicians believed this was the natural way to integrate functions.

Then Henri Lebesgue arrived and asked a revolutionary question:

Why are we slicing the x-axis at all?

This simple question led to one of the most important developments in modern mathematics.

The Riemann Perspective

Suppose we want to integrate the function

$$f(x)=x^2$$

on the interval

$$[0,1].$$

The Riemann approach divides the domain into many small intervals:

$$[0,0.1], [0.1,0.2], [0.2,0.3], \ldots$$

For each interval, we choose a height and construct a rectangle.

The area of each rectangle is approximately

$$\text{height} \times \text{width}.$$

Adding all these rectangles together gives an approximation to the area under the curve.

As the widths approach zero, the approximation converges to the true integral.

Conceptually, Riemann integration asks:

Where is the function?

Everything is organized according to the x-values.

A Different Way to Think

Lebesgue looked at the same problem and saw something different.

Instead of grouping points by their location on the x-axis, he grouped them according to the values taken by the function.

Rather than asking:

Where is x?

he asked:

What value does the function take?

This may sound like a small change, but it completely transformed integration.

An Everyday Analogy

Imagine you are studying a city with 1,000 residents.

One way to organize the data is by street address.

You might sort people by:

• House 1
• House 2
• House 3
• House 4

This is similar to the Riemann approach.

Now imagine organizing the same people by income.

You group them into:

• $0–$20,000
• $20,000–$40,000
• $40,000–$60,000
• $60,000–$80,000

This is similar to the Lebesgue approach.

The people have not changed.

Only the method of organization has changed.

Lebesgue realized that grouping according to values is often much more powerful.

Slicing the Y-Axis

For the function

$$f(x)=x$$

on

$$[0,1],$$

Riemann divides the x-axis:

$$0, 0.1, 0.2, 0.3, \ldots$$

Lebesgue effectively divides the y-axis:

$$0, 0.1, 0.2, 0.3, \ldots$$

and asks:

• How much of the domain produces values between 0 and 0.1?
• How much produces values between 0.1 and 0.2?
• How much produces values between 0.2 and 0.3?

Instead of measuring widths of intervals, he measures the size of sets where the function takes particular values.

The Mountain Analogy

Imagine a mountain landscape.

A Riemann-style survey divides the land into strips and measures the average elevation in each strip.

A Lebesgue-style survey divides the elevations into bands:

• 0–100 meters
• 100–200 meters
• 200–300 meters

and then asks:

How much land lies within each elevation band?

For example:

• 20 square kilometers between 0 and 100 meters
• 15 square kilometers between 100 and 200 meters
• 8 square kilometers between 200 and 300 meters

The total volume can then be computed by combining elevation levels with the amount of land occupying those levels.

This is the essence of Lebesgue integration.

Where Riemann Fails

The power of Lebesgue’s idea becomes clear when we consider functions that are too irregular for Riemann integration.

A famous example is the Dirichlet function:

$$
f(x)=
\begin{cases}
1, & x\in\mathbb{Q},\\
0, & x\notin\mathbb{Q}.
\end{cases}
$$

on the interval

$$[0,1].$$

This function equals 1 at every rational number and 0 at every irrational number.

Every interval contains both rational and irrational numbers.

As a result, Riemann’s rectangles never stabilize.

The Riemann integral does not exist.

Lebesgue’s Brilliant Solution

Lebesgue ignores the complicated arrangement of points and focuses on the values.

The function only takes two values:

$$0 \quad \text{and} \quad 1.$$

Now ask:

How much of the interval corresponds to the value 1?

The set is

$$\mathbb{Q}\cap[0,1].$$

This set has Lebesgue measure zero.

Therefore:

$$m(\mathbb{Q}\cap[0,1])=0.$$

How much corresponds to the value 0?

The irrational numbers occupy the entire interval except for a measure-zero set.

Therefore:

$$m([0,1]\setminus\mathbb{Q})=1.$$

The integral becomes

$$1\cdot0 + 0\cdot1 = 0.$$

The problem is solved immediately.

A function that defeated Riemann integration becomes completely manageable under Lebesgue integration.

The Fundamental Insight

Riemann integration asks:

1. Divide the domain.
2. Measure the function on each piece.
3. Add the contributions.

Lebesgue integration asks:

1. Group points according to function values.
2. Measure how much of the domain produces each value.
3. Add the contributions.

The difference may appear subtle, but it changes everything.

Why This Matters

Lebesgue’s idea became the foundation for much of modern mathematics.

It made possible:

• Modern probability theory
• Stochastic processes
• Functional analysis
• Fourier analysis
• Ergodic theory
• Quantum mechanics
• Statistical learning theory

Whenever mathematicians work with complicated functions, infinite-dimensional spaces, or probabilistic systems, they are almost always relying on Lebesgue’s framework.

Final Thoughts

The genius of Lebesgue was not that he invented a more complicated integral.

His genius was realizing that the problem should be viewed from a completely different angle.

Riemann asked:

Where is the function?

Lebesgue asked:

What values does the function take, and how much space corresponds to each value?

That change of perspective transformed integration from a geometric tool into one of the central ideas of modern mathematics.

More than a century later, it remains one of the most beautiful examples of how a simple shift in viewpoint can revolutionize an entire field.